Utility of the Ultrasound Examination of the Hand and Wrist Joints in the Management of Established Rheumatoid Arthritis

Authors


Abstract

Objective

To investigate the usefulness of point-of-care hand and wrist joint ultrasound (US) examination in patients with established rheumatoid arthritis (RA).

Methods

Fifty-one RA patients were evaluated using clinical disease activity measures and gray-scale and power Doppler (PD) US. Agreement between US and clinical findings and its impact on physicians' confidence and clinical decision were assessed.

Results

Agreement between intraarticular PD signal and joint swelling (JS) was moderate (82%; κ = 0.44). Agreement between PD signal and joint tenderness to palpation (TTP) was fair (75%; κ = 0.24). The greatest agreement between PD signal and clinical findings was seen in the 5th metacarpophalangeal (MCP) joint (96% JS, 88% TTP) and the poorest agreement was seen in the wrist (69% JS, 65% TTP) and 2nd (75% JS, 72% TTP) and 3rd (82% JS, 72% TTP) MCP joints. The presence of PD signal in nonswollen and/or nontender joints accounted for most of the disagreement in the wrists, while the opposite was true for the 2nd/3rd MCP joints. Agreement between sonographic synovial thickening and clinical findings was poor. Total sonographic synovial hypertrophy or PD score correlated significantly with physician-recorded, but not patient-recorded, clinical outcomes. US increased both physicians' confidence in their clinical decision (P < 0.0005, irrespective of Clinical Disease Activity Index score) and patients' confidence in physicians' medical decisions (88.4% of the cases). US modified biologic agent and/or disease-modifying antirheumatic drug (DMARD) use in 7 individual cases, but it did not affect the overall treatment plan (P > 0.15) or DMARD (P < 0.062) or biologic agent (P > 1.0) use in this group of RA patients.

Conclusion

PD examination of the wrist and 2nd/3rd MCP joints might be feasible and clinically meaningful in evaluation of disease activity in patients with established RA. US examination of the hand/wrist joints in RA increases physicians' confidence in their clinical decision and can help to individualize DMARD and biologic agent use.

INTRODUCTION

There is growing interest in the applications of musculoskeletal ultrasound (MSUS) in clinical rheumatology. Advances in ultrasound technology have enabled direct visualization of diverse pathologic features such as cortical bone erosions, synovial thickening, and synovial vascularity in the joints affected by rheumatoid arthritis (RA) with high sensitivity, specificity, and accuracy ([1-5]). MSUS may provide additional information about disease activity and response to therapy in RA patients. Several longitudinal studies have demonstrated that ultrasound examination can readily, and perhaps earlier than routine clinical outcome measures, detect response to therapy with biologic tumor necrosis factor (TNF) inhibitors ([6-8]).

Previous studies have shown that a significant proportion of RA patients classified as being in low disease activity or clinical remission while receiving conventional disease-modifying antirheumatic drugs (DMARDs) still have significant synovial hypertrophy and/or synovial hyperemia (specifically sonographic synovitis) measured by gray-scale (B-mode) and power Doppler (PD) ultrasound, respectively ([9, 10]). It seems that the addition of TNF inhibitors to DMARD therapy has led to a reduction of sonographic synovial hypertrophy in early RA; however, depending on the remission criteria used and the number of joints evaluated, 15–41.7% and 30.4–51% of patients with early and longstanding RA in clinical remission, respectively, still had detectable synovial PD signal even after introduction of a TNF inhibitor ([11-13]). As compared to synovial hypertrophy, measured on gray-scale ultrasound examination, the presence of PD sonographic synovitis was shown to be highly predictive of more frequent RA flares and worse radiographic outcomes ([10, 13-15]).

These observations have fostered a hypothesis that RA patients, who have synovial hyperemia/hypervascularity detectable by PD ultrasound, may benefit from a more aggressive or modified therapeutic intervention. Just as importantly, those who have no signs of sonographic synovitis, despite the presence of some subjective or even objective joint symptoms, should perhaps be treated less aggressively. Although direct data to support such a hypothesis are still lacking and the “acceptable degree” of sonographic synovitis has yet to be determined, a recent study showed that significant numbers of rheumatologists are either in favor of or already using joint ultrasound as part of the extended clinical evaluation in inflammatory arthritis at the individual patient level ([16]). In fact, the same study showed that in simulated cases, despite the lack of formal recommendations, rheumatologists were able to consistently integrate sonographic data in their clinical decision making and treatment recommendations. The present study was designed to further explore the potential utility of routine hand and wrist joint ultrasound in patients with established RA during a routine outpatient rheumatology office visit in a real clinical setting.

Box 1. Significance & Innovations

  • Clinical significance and practicalities of point-of-care ultrasound in routine assessment and management of rheumatoid arthritis (RA) patients are being actively debated. Our study is the first to demonstrate that data obtained during ultrasound examination of the hand and wrist joints in RA increase both physicians' confidence in clinical decision making and patients' confidence in physicians' medical decisions.
  • Prior studies have evaluated various approaches using limited joint ultrasound in assessment of RA activity, but have not provided any practically useful recommendations so far. We report that the largest disagreement between clinical and ultrasound findings exists in the wrist and 2nd and 3rd metacarpophalangeal joints, therefore making focused ultrasound examination of these joints clinically meaningful in routine assessment of disease activity in established RA.

PATIENTS AND METHODS

Patients

This study was approved by the University of California, San Diego Institutional Review Board/Human Research Protections Program. Fifty-one patients with established RA meeting the 1987 American College of Rheumatology criteria for RA ([17]) and 10 normal healthy control volunteers were enrolled in the study. Both the patients and physicians were consented for participation in the study. Patients with a history and/or suspicion of fibromyalgia were excluded.

Clinical evaluation

Four board-certified rheumatologists participated in recruitment and clinical evaluation of the patients. Patients documented self-recorded outcome measures on a standardized form that included detailed Health Assessment Questionnaire (HAQ) score, pain score (0–10 visual analog scale [VAS]), duration of the morning stiffness (minutes), fatigue (0–10 VAS), and patient's global assessment (0–10 VAS).

The physician evaluation included swollen joint count (SJC) and tender joint count (TJC; 28 joints counted, including the bilateral hand, wrist, elbow, shoulder, and knee joints) and physician's global assessment (PhG; 0–10 VAS). The RA Clinical Disease Activity Index (CDAI) was used as the composite disease activity measure owing to its immediate scoring and availability for clinical decision making (i.e., independent of the acute-phase reactants) ([18]).

Clinical decision making

After obtaining and reviewing the clinical data, the physicians were asked to indicate an overall likelihood of change in the management of the patient using the following quantitative Likert scale: 1 = very unlikely, 2 = unlikely, 3 = no change, 4 = likely, and 5 = very likely. Physicians were also asked to indicate specific changes they intended to make in the use of DMARDs, biologic agents, and systemic steroids. The following Likert scale was used to indicate the change: 1 = stop, 2 = decrease, 3 = no change, 4 = increase, and 5 = add/switch. Finally, the physicians were asked to indicate their level of confidence in their clinical decision on a VAS from 0–10, where 0 = a lack of confidence and 10 = the highest degree of confidence.

Patients then underwent sonographic examination of the hand and wrist joints by an experienced MSUS sonographer (AC) who was blinded to clinical details. The physicians were then provided with the standardized ultrasound report (see below), and were asked to make the clinical decision and indicate the level of confidence in that decision after they took into consideration the standardized hand/wrist ultrasound report. Physicians were also asked to indicate whether the ultrasound report was helpful for their overall clinical decision and/or future followup and management plans with the following answer choices included: very helpful, somewhat helpful, and not helpful.

Sonographic patient evaluation

A LOGIQ e ultrasound machine (GE Healthcare) equipped with a multifrequency 8–13-MHz linear transducer was used for MSUS. Ultrasound settings were standardized during the entire study, with B-mode frequency of 12–13 MHz, gain of 58–64%, PD frequency of 6.7 MHz, gain of 9–12%, and pulse repetition frequency of 0.6–0.8 Hz coupled with low-wall filters depending on the joint area and background interference. Gray-scale (B-mode) and PD examinations were performed in a standardized manner according to the European League Against Rheumatism guidelines ([19]). Ultrasound examinations were performed in a darkened room with an ambient room temperature after ≥20 minutes of settling in time. Dorsal and palmar joint regions and ulnar and radial aspects of the interphalangeal and second through fifth proximal interphalangeal (PIP) joints; dorsal and palmar joint regions of the first through fifth and ulnar and radial aspects of the second and fifth metacarpophalangeal (MCP) joints; and dorsal and volar joint regions and radial and ulnar aspects of the wrist (distal radioulnar, midcarpal, radiocarpal, and ulnar-carpal) joints were examined in long (sagittal plane) and short (axial plane) axes. Outcome Measures in Rheumatology definitions were used to assess joints for joint effusion, synovial hypertrophy, erosions, tenosynovial swelling, and hyperemia ([20]). Synovial hypertrophy (B-mode examination) was scored on a semiquantitative scale (where 0 = absence, 1 = mild, 2 = moderate, and 3 = severe) as described elsewhere ([21, 22]); the possible minimal–maximal total synovial hypertrophy score (tSH) was estimated between 0 and 135.

PD was performed on each of the hand/wrist joint regions and planes specified above. Synovial PD examination was graded on a semiquantitative scale (where grade 0 = no intraarticular color signal, grade 1 = single vessel signal[s], grade 2 = confluent color signal in less than half of the intraarticular area, and grade 3 = confluent color signal in more than half of the intraarticular area) as described in detail elsewhere ([22]); examples of PD synovitis grading for the MCP and wrist joints are shown in Figure 1. A possible total synovial PD score (tSPD) was estimated between 0 and 129. In addition, the presence or absence of sonographic effusion and erosions was documented for each joint. The standardized ultrasound report included tabulated descriptions of erosive changes, grading of synovial hypertrophy, and PD findings; presence or absence of joint effusion; and active pannus (cortical erosions with surrounding PD signal) in the individual PIP, MCP, and wrist joints. The second part of the report included a summary of the sonographers' impressions addressing the PIP, MCP, and wrist joints.

Figure 1.

Dorsal longitudinal power Doppler (PD) ultrasound scans of the metacarpophalangeal (MCP) and wrist joints show semiquantitative grading of PD synovitis in rheumatoid arthritis, where grade 0 = no intraarticular color signal, grade 1 = single vessel signal(s), grade 2 = confluent color signal in less than half of the intraarticular area, and grade 3 = confluent color signal in more than half of the intraarticular area. Radiocarpal (1) and midcarpal (2) joints were graded separately. The top 2 panels on the left-hand side show 2nd MCP joint scans and the bottom 2 panels on the left-hand side show 3rd MCP joint scans. Asterisks () indicate cortical bone erosions in the metacarpal heads. Arrows indicate PD signal in normal blood vessels (not a synovial blood flow). The top panel on the right-hand side shows arterial blood flow in a dorsal carpal arch and the second panel on the right-hand side shows venous blood flow in the superficial veins.

Exit questionnaire of patient satisfaction

After the sonographic examination, patients were provided with a standardized questionnaire and were asked to give their impressions on the 5 different statements (Table 1). Patients were asked to provide their impressions/answers on the following graded scale: strongly disagree, disagree, neutral, agree, or strongly agree.

Table 1. Rheumatoid arthritis patient opinion regarding the joint ultrasound examination performed during a routine clinic visit*
StatementYes, %No, %
  1. Results of the postultrasound/exit questionnaire.
20 to 30 minutes of ultrasound examination is a reasonable procedure time90.79.3
The information I have received about the results of the ultrasound was difficult to understand14.685.4
Ultrasound examination has increased my confidence in my physician's decisions about the management of my arthritis88.411.6
20 to 30 minutes of ultrasound examination is too long procedure time7.292.8
Ultrasound examination has increased my overall satisfaction with the care of my arthritis83.716.3
If given the option, I would be willing to have ultrasound of my joints done again92.87.2

Statistical analyses

Repeated-measures analyses of variance were used to test the overall effects of change in treatment before and after MSUS with followup t-tests for significant interactions; paired t-tests were used for comparison of confidence, treatment plan, and change in the specific medications before and after joint ultrasound in each case. Pearson's correlation coefficient was used for estimation of the relationship between 2 variables. Absolute agreements and unweighted kappa values between ultrasound and clinical examination were calculated, where kappa values <0.20 = poor, 0.21–0.40 = fair, 0.41–0.60 = moderate, 0.61–0.80 = good, and 0.81–1.00 = very good agreement. Kappa values represent a measure of by how much the observed agreement exceeds agreement by chance. Kappa values tend to be low if data are skewed, even if agreement is very good.

RESULTS

General characteristics of patients

Clinical characteristics of the RA study patients are shown in Table 2. Five patients were excluded from the final analysis because of missing data (2 patients) or disagreement between sonographic and clinical diagnosis (3 patients had sonographic features of calcium pyrophosphate dihydrate crystal deposition disease rather than erosive arthropathy). The mean age of the RA patients was 61.8 years (range 28–82 years) and the mean disease duration was 16.6 years (range 1–38 years). The majority of patients were receiving DMARDs, including methotrexate (60.9%; mean dosage 15.2 mg/week, range 5–25) and leflunomide (15.2%; mean dosage 17.5 mg/day, range 10–20) with or without combination therapy with a biologic agent (34.6%; either TNF inhibitor [n = 14], interleukin-6 inhibitor [n = 1], or rituximab [n = 1]). The mean age of the normal healthy control volunteers was 57 years (range 35–64 years).

Table 2. Demographic and clinical characteristics of the rheumatoid arthritis patients enrolled in the study*
 Value
  1. TNF = tumor necrosis factor; IL-6 = interleukin-6; HAQ = Health Assessment Questionnaire; VAS = visual analog scale (range 0–10); CDAI = Clinical Disease Activity Index.
  2. aPast treatment was defined as the treatment preceding 6 months before enrollment in the study.
  3. bJoint counts were done on 28 joints, including the bilateral hands, wrists, elbows, shoulders, and knees.
Sex, no. (%) 
Male4 (9)
Female42 (91)
Age, mean (range) years61.8 (28–82)
Disease duration, mean (range) years16.6 (1–38)
Hand/wrist radiograph in the past 2 years, no. (%) 
Erosions present15 (33)
Erosions absent13 (28)
Not done18 (39)
Treatment, no. (%) 
Methotrexate, past/activea8 (17.4)/28 (60.9)
Leflunomide, past/activea3 (6.5)/7 (15.2)
Sulfasalazine, past/activea8 (17.4)/0 (0)
Hydroxychloroquine, past/activea17 (37)/3 (6.5)
TNF inhibitor, past/activea14 (30.4)/14 (30.4)
IL-6 inhibitor, past/activea1 (2.2)/1 (2.2)
Rituximab, past/activea1 (2.3)/1 (2.2)
Prednisone within the past 3 months9 (19.6)
Patient-defined clinical characteristics 
HAQ score, mean (range)0.64 (0–2.5)
Pain score on VAS, mean (range)3.3 (0–10)
Patients with morning stiffness >30 minutes, no. (%)39 (84.8)
Fatigue on VAS, mean (range)4 (0–10)
Patient's global assessment on VAS, mean (range)3.3 (0–10)
Physician-defined clinical characteristics 
No. of tender joints, mean (range)b3 (0–11)
No. of swollen joints, mean (range)b2 (0–11)
Physician's global assessment on VAS, mean (range)2.6 (0–9.5)
CDAI score, mean (range)15.9 (0–50.6)
CDAI score <10, no.20
CDAI score 10–22, no.12
CDAI score >22, no.19

Agreement between clinical and sonographic joint examination

There was moderate agreement between clinically detected joint swelling and the presence of synovial PD signal in the wrist and the first through the fifth MCP joints (total agreement in all joints 82%; κ = 0.44), with the greatest agreement in the fifth MCP joint (96%) and the poorest agreement in the wrist (69%) and second (75%) and third (82%) MCP joints (Table 3). The disagreement was mainly due to the presence of synovial PD signal in the wrist joints, which clinically were categorized as nonswollen, while the somewhat greater disagreement in assessment in the second and third MCP joints was mainly due to the absence of detectable synovial PD signal in the joints that clinically were reported as swollen. Overall, agreement between the clinical joint tenderness and the presence of synovial PD signal was fair (total agreement in all joints 75%; κ = 0.24). There was significant disagreement between joint tenderness and the presence of synovial PD signal in the second and third MCP joints that was mainly due to the absence of PD signal in the second and third MCP joints that clinically were graded as tender, while a similar discrepancy in the wrist joints was mainly due to the presence of synovial PD signal in the joints that were graded as nontender (Table 3). Care was taken not to include vasculature of the dorsal carpal arch when scoring synovial PD signal in the wrist joints.

Table 3. Agreement between clinically detected JS (not graded) or TTP (not graded) and the presence/absence (not graded) of synovial PD signal in the wrist and MCP joints (left and right extremities combined) in patients with rheumatoid arthritis*
 WristMCP1MCP2MCP3MCP4MCP5
JSTTPJSTTPJSTTPJSTTPJSTTPJSTTP
  1. JS = joint swelling; TTP = tenderness to palpation; PD = power Doppler; MCP = metacarpophalangeal; PDUS+ = synovial PD ultrasound signal was detected, but clinical assessment reported no swelling or tenderness; PDUS− = synovial PD signal ultrasound was not detected, but clinical assessment reported presence of synovial swelling or tenderness.
Agree, frequency635277696760735877638572
PDUS+1928676112119927
PDUS−91261416191421318211
Agreement, %696587747572827287789688
Cohen's κ0.350.250.380.060.350.240.480.170.270.250.730.43

Agreement between sonographic synovial thickening and the clinically detectable joint swelling and/or tenderness was poor (47%; κ = 0.12 for swelling and 43%; κ = 0.04 for tenderness). The disagreement was mainly seen due to the presence of sonographic synovial thickening in the joints that clinically were classified as nontender and/or nonswollen (Table 4).

Table 4. Agreement between clinically detected JS (not graded) or TTP (not graded) and the presence/absence (not graded) of GSUS synovial hypertrophy in the wrist and MCP joints (left and right extremity joints combined) in patients with rheumatoid arthritis*
 WristMCP1MCP2MCP3MCP4MCP5
JSTTPJSTTPJSTTPJSTTPJSTTPJSTTP
  1. JS = joint swelling; TTP = tenderness to palpation; GSUS = gray-scale ultrasound; MCP = metacarpophalangeal; GSUS+ = synovial hypertrophy reported on ultrasound, but clinical assessment reported absence of synovial swelling or tenderness; GSUS− = synovial hypertrophy was not detected on ultrasound, but clinical assessment reported presence of synovial swelling or tenderness.
Agree, frequency414143443530483545453935
GSUS+454744405255364343365050
GSUS−5325245111804
Agreement, %454548492634543951514439
Cohen's κ0.070.110.100.090.060.010.170.030.070.050.100.01

Agreement between the presence of synovial PD signal and clinical assessment of joint swelling was moderate in patients with high disease activity (CDAI score >22: 78%; κ = 0.51), fair in patients with low disease activity (CDAI score <10: 86%; κ = 0.36), and poor in patients with moderate disease activity (CDAI score 10–22: 81%; κ = 0.20). Agreement between the presence of ultrasound PD signal and clinical assessment of joint tenderness was 84%, 79%, and 59% for low (CDAI score <10), moderate (CDAI score 10–22), and high disease severity (CDAI score >22), respectively. Cohen's kappa coefficients were near zero for these measures because biases in the data resulted in elevated chance probability. Disagreement in low-severity patients was largely due to clinical report of symptom absence and PD report of signal presence, while the reverse was true in high-severity patients.

Correlation between hand joint ultrasound findings and patient- and physician-recorded clinical measures

Clinical characteristics of the RA patients included are shown in Table 2. Physician examination reported a mean TJC of 3 (range 0–11), a mean SJC of 2 (range 0–11), a mean PhG score of 2.6 (range 0–9.5), and a mean CDAI score of 15.9 (range 0–50.6). Sonography showed on average 9 joints with erosions (range 0–19), a mean tSH of 37 (range 4–132), and a mean tSPD of 6.7 (range 0–35). Examination of 10 healthy volunteers revealed 1 case with a small lunate erosion (cortical break of >1 mm), 2 cases with low-grade synovial hypertrophy at least in one view of 1 wrist and 1 second MCP joint; one of the subjects had a grade 1 PD signal that could have been attributable to a penetrating cortical blood vessel rather than a synovial signal. Sonographic examination of the PIP joints revealed a high prevalence of osteophytosis/degenerative changes in both the RA and healthy volunteer groups. Confounding by significant osteoarthritis of the PIP joints was viewed as a potential source of interference with clinical joint assessment, particularly with assessment of joint tenderness. PIP joints were, therefore, excluded from clinical–sonographic agreement analysis.

There was no significant correlation between patient-recorded intensity of pain, HAQ score, or patient's global assessment score and any of the ultrasound measures (Table 5). There was a significant correlation between physician-recorded clinical outcomes and hand joint ultrasound (Table 5). These included a significant correlation between the presence of sonographic erosions and SJC (r[44] = 0.2543, P < 0.05), but the presence of sonographic erosions did not correlate with TJC, PhG score, or CDAI score. There was a significant correlation between tSH and TJC (r[44] = 0.4314, P < 0.005), tSH and SJC (r[44] = 0.4804, P < 0.005), tSH and PhG score (r[44] = 0.2447, P < 0.05), and tSH and CDAI score (r[44] = 0.4612, P < 0.005). tSPD correlated significantly with TJC (r[44] = 0.3559, P < 0.01), SJC (r[44] = 0.6548, P < 0.005), PhG score (r[44] = 0.4498, P = 0.005), and CDAI score (r[44] = 0.5043, P < 0.005). The presence of sonographic joint effusion did not correlate with PhG score, but did correlate with TJC (r[44] = 0.3449, P < 0.01), SJC (r[44] = 0.3270, P < 0.05), and CDAI score (r[44] = 0.3627, P < 0.01).

Table 5. Correlations (r) between hand and wrist joint ultrasound and patient- and physician-recorded clinical outcomes in established rheumatoid arthritis patients*
US measuresPatient-recorded outcomesPhysician-recorded outcomesDisease activity, CDAI
PainHAQGlobal scoreTJCSJCGlobal score
  1. US = ultrasound; HAQ = Health Assessment Questionnaire; TJC = tender joint count; SJC = swollen joint count; CDAI = Clinical Disease Activity Index; tSH = total synovial hypertrophy score estimated by B-mode US; tSPD = total synovial power Doppler score estimated by power Doppler US.
  2. aP < 0.05.
  3. bP < 0.005.
  4. cP < 0.01.
Sonographic erosion0.12170.0235−0.03090.04900.2543a−0.00450.1167
tSH0.2227−0.08050.02820.4314b0.4804b0.2447a0.4612b
tSPD0.1170−0.05630.00790.3559c0.6548b0.4498b0.5043b
Sonographic effusion0.2093−0.03450.05180.3449c0.3270a0.22770.3627c

Impact of the hand and wrist joint examination on the management of patients with established and currently treated RA

There was an overall significant increase in physician-reported confidence in clinical decision making after joint ultrasound data were available (mean confidence rating of 7 before ultrasound and 8.1 after ultrasound; P < 0.0005) (Table 6). Overall, physicians reported intention of treatment plan change in 14 of 46 cases based on routine examination/assessment alone, and in 16 of 46 cases after the ultrasound report was available. Of the latter 16 cases, management was directly affected by the ultrasound in 7 cases (the change in both intention and use of specific medication occurred in these cases). In these cases, 2 of 7 were due to reversal, 4 of 7 were due to new intention, and 1 of 7 was due to switch from initial intention to use of a biologic agent; 3 of 7 cases were pertaining to the use of DMARDs and 4 of 7 to the use of a biologic agent. These changes on the individual patient level were not sufficient to produce a statistically significant effect on the overall physician-reported intention of treatment plan change (P > 0.15) or use of a biologic agent (P > 1.0), DMARD (P < 0.062), or steroid (P > 1.0) (Table 6) when management intentions were reported on a Likert scale.

Table 6. Impact of the sonographic hand and wrist joint examination on physicians' clinical confidence and management decisions in established rheumatoid arthritis patients*
 t-testP
  1. DMARD = disease-modifying antirheumatic drug.
  2. aRecorded on a 0–10 visual analog scale.
  3. bRecorded on the following scale: 1 = very unlikely, 2 = unlikely, 3 = no change, 4 = likely, and 5 = very likely.
  4. cRecorded on the following scale: 1 = stop, 2 = decrease, 3 = no change, 4 = increase, and 5 = add/switch.
Physician confidence in clinical decisionat(44) = −3.85< 0.0005
Intention to change treatment planbt(45) = −1.48> 0.15
Change in DMARD usect(45) = 1.91< 0.062
Change in biologic agent usect(45) = 0.00> 1.0
Change in steroid usect(45) = 0.00> 1.0

Data stratification based on the CDAI score showed significant intent to change treatment plan in patients with more severe compared to less severe disease activity both before (F[2,43] = 7.4, P < 0.002) and after (F[2,43] = 3.97, P < 0.03) ultrasound. There was a significant increase in physicians' confidence of clinical decision after ultrasound regardless of disease activity (F[1,42] = 13.94, P < 0.006). Physicians reported that they were more likely to increase the use of DMARDs in established RA patients with a higher CDAI score based on clinical examination alone (F[2,43] = 3.55, P < 0.0374), but following ultrasound, there was a reversal of such intent to change DMARD use in these patients (F[2,43] = 3.64, P < 0.035). Physicians indicated that the ultrasound report was very helpful in their overall clinical decision and/or future followup and management in 14 (30%) of 46 cases. The mean CDAI score in these cases was 14.3 (range 3.3–26.6).

Patient satisfaction

The results of the patient satisfaction/exit questionnaire are shown in Table 1; 90.7% of patients thought that 20–30 minutes is a reasonable procedure time, 92.8% of patients would be willing to have another ultrasound examination if recommended, 88.4% of patients reported that they have more confidence in their physician's management decisions, and 83.7% stated that the hand and wrist joint ultrasound had improved their overall satisfaction with care. A minority of patients (14.6%) reported that the joint ultrasound results were difficult to understand.

DISCUSSION

Despite all of the accumulated evidence to support the potential role of joint ultrasound utility in daily clinical practice in the management of RA, there are only a few studies that indirectly addressed this question ([23-25]). Therefore, the purpose of our study was to investigate whether performing point-of-care ultrasound examination of the hand and wrist joints in patients with established RA may provide clinical data useful in assessment of disease activity and how it would impact clinical decision making and use of DMARDs, biologic agents, and steroid medications during routine outpatient rheumatology office visits.

In exploring the potential value of point-of-care joint ultrasound examination in assessment of disease activity in RA, we have directly compared agreement between clinical and sonographic joint examination, rather than overall disease activity and sonographic scores of synovitis. It appears that most of the discrepancy between clinical examination and synovial PD signal was found in the wrist and second and third MCP joints. The presence of synovial PD signal in the joints that clinically were categorized as nontender and/or nonswollen accounted for most of the discrepancy in the wrists. In contrast, most of the discrepancy between clinical and sonographic synovitis in the second/third MCP joints was accounted for by the cases that had no detectable synovial PD signal despite being categorized as tender and/or swollen. This suggests that in contrast to missing PD sonographic synovitis in the wrist joints, physical examination perhaps tends to overestimate the presence of active synovitis in the second and third MCP joints if a gold standard for detection of synovitis is defined by the presence of synovial PD signal. The opposite might be true in some cases, but it seems to be much less common based on our data. Our findings in the MCP joints are also supported by the previous magnetic resonance imaging–controlled studies in which physical examination of the MCP joints was reported to have only a weak correlation with PD ultrasound ([2]), and physical examination reportedly found swelling of the second/third MCP joints more commonly than ultrasound synovitis ([21]).

Prior studies have explored the potential value of limited joint ultrasound as a possible substitute for 28- or 44-joint evaluations, the approach that would be much more feasible in real clinical practice ([26, 27]). In RA patients who were in disease remission defined by a Disease Activity Score in 28 joints <2.6 or a Simplified Disease Activity Index score <3.3, the number of joints with synovial PD signal was shown to be almost identical irrespective of whether 12 joints, 7 joints, 6 joints, hand/wrist, or reduced hand (excludes PIP joints)/wrist joints and, therefore, regardless of whether the feet and PIP joints were included in the evaluation. Our patient population was similar in demographic characteristics, but had overall higher disease activity based on the CDAI score. Based on our data, it seems that in established RA, sonographic evaluation of the fourth and fifth MCP joints adds little to the physical examination and that ultrasound of the PIP joint shows a very high prevalence of degenerative joint disease, and therefore is less useful in evaluation of disease activity in RA. Our study results support the notion that targeted PD sonography of the bilateral wrist and second and third MCP joints could be feasible and clinically meaningful for evaluation of RA patients with established disease.

We have explored agreement between clinical synovitis determined by the presence of tenderness and/or palpable joint swelling versus gray-scale synovitis defined by the presence of sonographically detectable thickening of synovial space. We found a much larger (relative to PD data) discrepancy between sonographic synovial hypertrophy and clinically recorded joint swelling and/or tenderness in all of the MCP and wrist joints. In contrast to PD findings in this respect, the discrepancy between clinical synovitis and gray-scale synovial hypertrophy was more universal and was found at approximately the same percentage range across all of the MCP and wrist joints. This discrepancy seems to have been accounted for primarily by the presence of sonographically detectable synovial thickening in the MCP and wrist joints that clinically did not feel tender and/or swollen. In addition to operator dependence, there are several inherited methodologic, pathologic, and anatomic factors that may account for this relatively large discrepancy. Some of the potential reasons for somewhat low reliability of the sonographic synovial hypertrophy measurement were recently addressed in an expert review ([28]). Most pertinent to our study would be the older age of the patients and the presence of degenerative disease in the PIP joints, both of which make measurement of sonographic synovial hypertrophy less reliable ([28]). In addition, the originally proposed semiquantitative assessment of synovial hypertrophy has been based on the hypoechoic/anechoic synovial space elevation above the proximal phalanx relative to the MCP joint capsule on a volar view. Although universally accepted, the same approach is likely less applicable for wrist joints and to dorsal views of the MCP joints, especially for when low-grade synovial hypertrophy is present. In addition, gray-scale ultrasound examination very likely tends to overestimate synovial proliferation in its clinicopathologic terms in established and treated RA, in which sonographic synovial thickening maybe overrepresented by “inactive” tissue rather than active proliferation. Therefore, our data indicate that despite overall good correlation of the tSH with physician-recorded clinical outcome measures (TJC, SJC) and the CDAI, the sonographic measurement of MCP and wrist joint synovial hypertrophy during point-of-care ultrasound examination is of less practical significance (than PD). This also can be supported by the fact that the correlation between the tSH and PhG scores was only modest compared to the very high correlation between the tSPD and PhG scores.

At present, there are no outcome data–driven recommendations as to what is considered to be “treatable” sonographic synovitis, or how big the discrepancy between clinical and sonographic assessment should be to trigger therapeutic intervention, e.g., either increase, change, or back off therapy with a DMARD and/or a biologic agent in RA. With that in mind, the goal of our study was to explore the potential impact of point-of-care ultrasound examination on physicians' clinical decision making, including confidence, intention of treatment change, and use of specific RA medications during a routine followup visit. The results of our study showed that availability of the ultrasound report resulted in a significant increase in overall physician confidence in their clinical decision making irrespective of the RA disease activity measured by the CDAI. As a result, hand and wrist ultrasound report data led to bidirectional changes in management of individual cases. Overall, however, when reported on a quantitative Likert scale, ultrasound data did not change physicians' intentions of treatment plan or use of biologic agents or steroids in this RA patient population with established and treated disease. Notably, availability of the ultrasound data has reversed initial physicians' intention to increase/escalate DMARD use in patients with more severe disease as measured by the CDAI. This finding is somewhat counterintuitive. The explanation for it comes from the notion of the high prevalence of degenerative disease and, therefore, significant tenderness observed in the PIP joints in our study population. This has translated into higher CDAI scores and physicians' intention to escalate DMARD therapy based on physical examination alone. Several other factors should be taken into account while interpreting the results of physician clinical decision. First, our study was likely underpowered to reach a more definite conclusion in relation to disease activity, duration, prior response, or current use of different medications because it included RA patients with established and treated disease. The study was also confounded by the lack of standardized recommendations of joint ultrasound use in the management of RA. Therefore, it is likely that different physicians put different weight into the joint ultrasound data while reevaluating their intentions to change treatment plan or use of DMARDs and biologic agents after reviewing the ultrasound report. Such uncertainty of how to use additional clinical data obtained by the ultrasound examination exists in the rheumatology community, and our data showed how this translates into a lack of a more defined (i.e., beyond the individual cases) ultrasound data impact on the management of RA. That being said, the lack of overall significant change in intention to use biologic agents is an important observation. Because of the high prevalence of sonographic synovitis (often viewed as subclinical synovitis) in patients who are receiving DMARDs alone or even receiving combination therapy with a DMARD/biologic agent ([9, 11]), there is concern of a possible significant increase in biologic agent use in routine management of RA if decisions are based on joint ultrasound results. The data obtained in this study do not support that.

The impact of the ultrasound examination on patient satisfaction with their care by rheumatologists has not been systematically studied. Nonetheless, a recent pilot study from Australia showed that patients with early RA who underwent sonographic visualization of their joints reported greater perceived understanding and a trend toward increased compliance with the management ([29]). Our postultrasound patient satisfaction exit questionnaire clearly demonstrated that the procedure was well received by the patients, and 88.4% of the RA patients stated that ultrasound examination increased their confidence in physicians' recommendations. This is a significant aspect of the potential impact of ultrasound on the management of RA patients. The point might be demonstrated by the fact that in our study, at least 3 patients (6%) felt very strongly they had active RA, and therefore requested additional intervention, but easily accepted physicians' recommendations not to escalate RA therapy after ultrasound confirmed initial physicians' assessment that did not suggest clinical synovitis.

In summary, our study results suggest that targeted PD examination of the wrist and second/third MCP joints may be clinically meaningful and feasible during routine assessment of RA patients. Based on current understanding of the significance of ultrasound findings in RA joints, ultrasound of the hand and wrist joints may increase physicians' confidence in their clinical decision and can help to make individual treatment decisions.

AUTHOR CONTRIBUTIONS

All authors were involved in drafting the article or revising it critically for important intellectual content, and all authors approved the final version to be published. Dr. Ceponis had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.

Study conception and design. Ceponis, Kavanaugh.

Acquisition of data. Ceponis, Onishi, Bluestein, Kalunian, Kavanaugh.

Analysis and interpretation of data. Ceponis, Townsend, Kavanaugh.

Ancillary